The lights dim at the Taipei Nangang Exhibition Center. It's May 31, 2026 — the opening keynote of Computex — and Jensen Huang, leather jacket and all, stands beneath a screen displaying a single phrase: "The PC is being reinvented." Over 30 years, NVIDIA transformed from a graphics card company into the world's most important AI infrastructure provider. But this moment is different. Jensen isn't talking about data centers or server racks. He's talking about the machine on your desk — and the laptop in your bag.
"For forty years, you launched apps. Click. Type," Huang tells the crowd. "With RTX Spark and Microsoft Windows, you ask — and the PC does the work."
And with that, NVIDIA RTX Spark was born.
The Personal AI Computer Is Here
RTX Spark isn't just another processor. It's a full-system-on-a-chip that brings together a 20-core NVIDIA Grace Arm CPU with a Blackwell RTX GPU — connected via NVLink-C2C — delivering up to 1 petaflop of FP4 AI performance and support for up to 128GB of unified memory. MediaTek collaborated on the custom CPU design, contributing to what NVIDIA calls "best-in-class power efficiency." The chip fits into slim Windows laptops with all-day battery life and compact desktop PCs.
In plain English? This thing can:
- Run a 120-billion-parameter LLM with up to 1 million tokens of context — locally, on-device, no internet required
- Generate 4K AI video, render 90GB+ 3D scenes, and edit 12K 4:2:2 video
- Play AAA games at 1440p at over 100 FPS, with DLSS, Reflex, and G-SYNC all baked in
Adobe is so confident in RTX Spark that it's rearchitecting Photoshop and Premiere Pro from the ground up for the platform, promising 2x faster AI and graphics performance.
Where AI Agents Finally Come Home
The real headline, though, isn't the silicon — it's what the silicon enables. AI agents like OpenClaw and Hermes Agent have hit an inflection point on developer platforms like GitHub and OpenRouter, but broad adoption has been blocked by one fundamental problem: security and privacy. Users want agents that work locally, not in some data center they don't control.
NVIDIA and Microsoft are solving this together. The collaboration introduces new Windows security primitives — identity, containment, policy, and end-to-end security — paired with NVIDIA OpenShell, a runtime that lets users define exactly what agents can and cannot do. It can intelligently route queries to local vs. cloud models based on privacy policies, and even disguise personal information before it leaves the device.
"At Nous, we expect tasks to increasingly run on device as personal agents like our Hermes Agent become more capable and ubiquitous," said Dillon Rolnick, CEO of Nous Research. "You realize you're buying a full-fledged assistant, not a typical laptop."
The first Windows-native agent apps — including Hermes Agent and OpenClaw — are already being built for the platform. They'll be able to execute tasks across Windows applications, reason through cross-app workflows, generate images and video, code plugins, and semantically search local files. All under your control. All private.
Coming This Fall — But Start Saving
RTX Spark-powered laptops and compact desktops will arrive in fall 2026 from the biggest names in the industry:
- ASUS, Dell, HP, Lenovo, Microsoft Surface, and MSI (launch partners)
- Acer and GIGABYTE (to follow shortly after)
- Microsoft's own Surface Laptop Ultra and Surface RTX Spark Dev Box lead the charge
But premium performance commands a premium price. According to Morgan Stanley analysts who surveyed vendors at Computex, RTX Spark laptops will likely start around $1,799 for the N1 tier (8+4 core configuration) and climb to $2,899+ for the N1X high-end models (10+10 core, 48 streaming multiprocessors). These aren't budget machines — they're workstations wrapped in laptop form factors.
What This Means
Three takeaways as we digest the biggest PC announcement in years:
- The PC platform war just got real. Arm-based Windows, backed by NVIDIA's full CUDA and RTX ecosystem, signals a direct challenge to both Intel/AMD x86 dominance and Apple's M-series silicon. For the first time, Windows on Arm has a serious alternative to Qualcomm.
- Local AI is no longer a compromise. Running a 120B-parameter model locally with 1M-token context wasn't a realistic laptop use case — until now. RTX Spark flips the narrative that you "need the cloud" for real AI work.
- Your next computer will be an assistant. Jensen said it best: the PC is moving from tool to teammate. The combination of on-device AI performance and secure agent runtimes means your computer won't just run programs — it will act on your behalf.
The PC has had a few defining moments: the IBM PC in 1981, the graphical interface in 1984, the internet in the 1990s, the smartphone in 2007. RTX Spark might just be the next one. It's not a faster computer — it's a different kind of computer. One that listens, reasons, and acts. One that moves from tool to teammate.
And it arrives this fall.
Comments